WO2008110704A1

WO2008110704A1 - Method for balancing the movement of mobile masses in a bi-linear electrodynamic motor

Info

Publication number: WO2008110704A1
Application number: PCT/FR2008/050053
Authority: WO
Inventors: Jonathan Buquet; Gérald AIGOUY; Thierry Trollier
Original assignee: L'Air Liquide Société Anonyme pour l'Etude et l'Exploitation des Procédés Georges Claude
Priority date: 2007-03-14
Filing date: 2008-01-11
Publication date: 2008-09-18
Also published as: ES2371724T3; EP2137588A1; FR2913782A1; US20120056565A1; US8749112B2; EP2137588B1; ATE522854T1

Abstract

The invention relates to a method for balancing the movement of magnetised mobile masses (10, 20) in a bi-linear electrodynamic motor that comprises two mobile masses (0, 20) moving in opposite directions parallel to the axis (x-x) of the motor, characterised in that said method comprises the following steps: providing at least one first magnetic sensor (12) and at least one second magnetic sensor (22) capable of respectively supplying a first electric signal (s1(t)) and a second electric signal (s2(t)) respectively representative of the movement of the first (10) and second (20) mobile masses; storing a first error signal ( Δs(t) ) equal to a difference between said first (s1(t)) and second (s2(t)) electric signals and carrying out a harmonic analysis of said error signal; applying to the first mobile mass (10) a sinusoidal excitation signal ( e1(t) ) at a first given frequency (f0); iteratively applying to the second mobile mass N successive excitation signals ( e2n (t)), an excitation signal of a rank n (O ≤ n ≤ N - 1) being equal to a n-order Fourier series at a fundamental frequency f0, the n-order term of said series being phase- and amplitude-adjusted in order to minimise the component at the n. f0 frequency of said error signal; and taking as an excitation signal of the second mobile mass the excitation signal ( e2n-1 (t) ) having a rank N - I. The invention can be used for reducing the vibrations of bi-linear electrodynamic motors.

Description

Method for balancing the movement of moving masses of a bilinear electrodynamic motor

The present invention relates to a method for balancing the movement of moving masses of a bilinear electrodynamic motor.

The invention finds a particularly advantageous application in the field of cryogenic machines with alternating cycle, Stirling machines or pulsed gas tubes, implementing bilinear electrodynamic motors with moving masses forming pistons, and more particularly cryogenic machines intended to be loaded into spacecraft such as Earth observation satellites. In this application, bilinear electrodynamic motors act as compressors of the fluid used, for example helium. The principle of operation of a bilinear electrodynamic motor is based on the generation by induction coils of cyclic magnetic forces that animate a rectilinear motion the magnetic moving masses constituting the engine pistons, which are mounted on mechanical bearings which, due to their construction, develop an axial elastic return force proportional to the displacement of the moving masses. The latter are therefore characterized by a mechanical resonance frequency determined by the mass in motion, the stiffness of the bearing, the magnetic stiffness and the fluidic load.

The control of the motor then consists in applying to the induction coils an excitation current at the mechanical resonance frequency of the magnetic moving masses, so as to obtain a natural amplification of the movement of movement of the pistons.

In bilinear electrodynamic compressors, the moving masses of the pistons are aligned in the same compression chamber and oscillate in mechanical opposition to the frequency of the excitation current of the coils, generally a sinusoidal current. This assembly has the advantage of a natural balance moving moving masses, which is not the case of linear compressors mono-pistons. However, the tolerances on the mechanical and magnetic parameters, such as mass, mechanical and magnetic stiffness, alignment, etc., lead to mechanical responses of the two half-motors slightly different for an identical electrical instruction, and therefore induce vibrations of the motor in the axis of displacement of the movable masses of the pistons.

In an application to the observation of the Earth by satellite, this residual vibratory level leads to the degradation of the shooting, this all the more as the severe mechanical environment during the launch in terms of vibrations and shocks of the launcher , as well as the thermal environment in orbit excluding any heat transfer by convection, impose to fix the compressor rigidly on the structure of the satellite, which promotes the propagation of the vibrations generated by the compressor towards the other equipment also fixed on the structure satellite, including cameras.

In addition, given the lifetimes required, between 5 and 10 years, it is necessary to follow the evolution of the balancing of the compressor to ensure a minimum level of vibrations induced throughout aging. Current solutions for reducing the residual vibrations due to a defective balancing of the movement of the moving masses consist in optimizing the setpoint of the driving current of one of the moving masses relative to the other, according to a master-slave system.

For this purpose, force sensors or accelerometers are placed in mechanical relation with the compressor so as to provide a measurement of the vibrations induced on the compressor by a possible imbalance between the displacements of the two pistons. The optimal control current of the piston-slave is obtained when the vibration measurement collected by the force sensors or accelerometers is minimal. The force sensors are piezoelectric washers placed at the compressor mechanical interfaces with the satellite structure. This type of sensor, however, has a number of disadvantages. First, if they are able to measure the residual vibrations specific to the compressor, the force sensors can also record those from other equipment attached to the same mechanical structure of the satellite. The measurement of the desired vibrations is thus disturbed by the mechanical environment of the compressor.

The piezoelectric sensors being of poor thermal conductors, it is necessary to provide a thermal path different from the mechanical path passing through the washers to evacuate the heat dissipations of the compressor, namely the heat of compression of the gas, the losses by Joule effect, by current of Foucault, by hysteresis, etc. For example, an ad hoc thermal path can be achieved by conductive braids placed in short circuit on the piezoelectric washers. This obviously results in a complex and more expensive integration.

Finally, it is very difficult to obtain a redundancy of these force sensors, given their specific mechanical layout.

Similarly, the use of accelerometers arranged on the compressor does not lead to satisfactory results for the following reasons.

The measurement provided by the accelerometers generally has a low signal-to-noise ratio because of the large masses on which the compressor is fixed. In addition, the transmitted force is reconstituted by interpreting the acceleration measurement according to an effective mass, resulting from the moving masses, which is difficult to evaluate and therefore imprecise.

As for the force sensors, the acceleration measurement is disturbed by the mechanical environment around the compressor, which leads to the measurement of accelerations that do not depend on the compressor.

In reality, the acceleration measurement is well suited to mounting the compressor in suspension and not to mounting on a rigid interface by bolting.

However, a traditional suspension installation, necessary for a correct measurement of the acceleration, decouples the structure of the interfaces and therefore imposes conditions that are difficult to adapt to spatial applications, such as the creation of a specific thermal path to evacuate the heat dissipations. and the establishment of a blocking mechanism the suspension, then unblocking when the compressor has to withstand external mechanical loads.

Finally, accelerometers and their associated packaging are expensive.

Also, the object of the invention is to propose a method of balancing the movement of the moving magnetic masses of a bilinear electrodynamic motor, which would allow the implementation of the master-slave control system mentioned above, from measurements of displacement of moving masses that would not be disturbed by the mechanical environment external to the engine. This object is achieved, according to the invention, because said method comprises steps of:

providing at least one first magnetic sensor and at least one second magnetic sensor able respectively to deliver a first electrical signal and a second electrical signal respectively representative of the movement of a first and a second mobile mass,

recording an error signal equal to the difference between said first and second electrical signals and performing a harmonic analysis of said error signal,

applying to the first mobile mass a sinusoidal excitation signal at a given frequency / ₀ ,

- iteratively apply to the second moving mass N successive excitation signals, an excitation signal of rank n (O ≤ n ≤ NY) being equal to a Fourier series of order n of fundamental frequency / ₀ , the term d order n of said series being adjusted in amplitude and in phase so as to minimize the component at the frequency n. f _{0 of} said error signal,

take as the excitation signal of the second mobile mass the excitation signal obtained at the iteration of rank N-1.

Thus, it is firstly understood that the method according to the invention operates on the basis of signals representative of the displacement of the moving masses provided by magnetic sensors, such as Hall effect sensors placed for example on the motor housing, intercepting a density, variable according to their displacement, the magnetic flux generated by the magnetic moving masses. As a result, the displacement measurements thus obtained are independent of the engine environment, at least as long as no other equipment in the vicinity of the engine provides a variable magnetic flux density. Furthermore, it is also understood that the iterative method proposed by the invention consists in constructing an excitation signal for the mobile-slave mass as a Fourrier series, each term of which minimizes the corresponding harmonic of the error signal, the signal applied to the mobile-master mass being the sinusoidal excitation signal at the fundamental frequency, applied to the induction coil associated with this mobile-master mass. Under these conditions, optimum balancing of the movement of the moving masses is achieved and, consequently, a reduction in the minimum residual vibrations.

The invention also has many other advantages. The balancing of the moving masses can be carried out at any time, in particular during the duration of the mission of the satellite in order to take account of the aging of the organs of the engine.

Unlike known measurement systems based on force sensors or accelerometers, no particular modification or adaptation must be made to the mechanical and thermal interfaces of the motor with its environment, such as for example a compressor fixed to the structure of a motor. satellite.

The signals representative of the movement of the moving masses are obtained without resorting to intrusive sensors which would affect the operation of the engine.

Finally, it is very easy to perform a redundancy of the system by placing several magnetic sensors in various places of the motor housing, the exact location of the sensors being of no importance from the moment it is able to intercept a sensor. magnetic flux density sufficient to obtain a minimum signal-to-noise ratio.

The following description with reference to the accompanying drawing, given by way of non-limiting example, will make it clear what the invention is and how it can be achieved. Figure 1 is a sectional view of a compressor equipped with magnetic sensors for implementing the method according to the invention.

FIG. 1 shows a linear electrodynamic motor intended, for example, to be integrated as a compressor with a Stirling-type alternating-cycle cryogenic machine, a pulsed gas tube or the like, or Joule-Thomson type continuous-flow cryogenic machines. . The motor of FIG. 1 comprises two moving masses 10, 20 forming two pistons charged with compressing a cryogenic fluid, such as helium.

In operation, the two mobile masses 10, 20 move in opposite directions parallel to the axis XX of the motor in an alternating movement whose frequency / ₀ is chosen substantially equal to the resonant frequency of the electromotor assembly and its load fluidics. A typical value of this resonance frequency is for example 50 Hz. At this frequency, the amplitude of the movement of the moving masses is then maximum and is limited only by the damping forces due to the various mechanical friction, which are as low as possible for maximum motor efficiency.

Moreover, the reciprocating movement of the pistons is obtained by applying to induction coils 11, 21 a sinusoidal excitation signal at the frequency / ₀ . The magnetic coupling of the pistons with the coils 11, 21 is achieved by means of permanent magnets carried by the moving masses 10, 20.

As mentioned above, in spite of the fact that the bilinear motor of FIG. 1 is designed so that the movement of the moving masses is naturally balanced, it can occur for various reasons of slight imbalances in amplitude and phase between the displacements of the moving masses 10, 20, with the consequence of the appearance of residual vibrations responsible for degradation in the quality of the shooting of the satellite cameras. To limit these induced vibrations, it is intended to equip the motor with magnetic sensors 12, 22, Hall effect sensors for example, able to respectively provide a first electrical signal .S ₁ (t) representative of the movement of the mass 10 and a second electrical signal ^ ₂ (t) Representative of the movement of the mass 20. These electrical signals .S ₁ (t) and .S ₂ (t) originate from the variation of the magnetic induction through the sensors 12, 22, due to the variation of the density of the magnetic flux created by the magnetic masses during their movement, as shown in Figure 1 by divergent magnetic field lines from the magnetized masses 10, 20. During the movement of the moving masses, the sensors intercept more or less of field lines, hence the variation of magnetic flux and the induced current resulting therefrom.

In the example of FIG. 1, the magnetic sensors 12, 22 have been placed on the longitudinal axis X-X of the motor. Of course, they could be placed at another place on the motor housing, for example laterally, the only condition being that they can detect variations in the density of the magnetic flux created by the moving magnetic masses 10, 20.

The movement of the moving masses 10, 20 is obtained by applying to the coil 11 of the first mobile mass 10, which will later be chosen as the master mass, an excitation signal (), and to the coil (21). the second mobile mass 20, which will be chosen as ground-slave, an excitation signal e ₂ (t). These excitation signals are periodic of frequency / ₀ .

If the two half-motors are perfectly balanced, the difference Δs (t) = .S ₁ (t) -s ₂ (t), which will be called error signal, is zero. However, we have seen above that in practice there exists between the two movements an imbalance, source of residual vibrations, that the invention seeks to correct at best.

For this, the error signal Δs (t) of periodic frequency / ₀ is recorded and subjected to a harmonic analysis so as to perform a decomposition into N Fourier components of frequency nf ₀ with

0 ≤ n ≤ N-1, where N is an arbitrary given number chosen according to the level of correction sought.

The respective amplitude of the Fourier components of the error signal Δs (t) will be denoted by C ₀ , ..., C _n , ..., C _N-1 .

A sinusoidal excitation signal at the frequency / ₀ is applied to the master mass 10: eβ) = A _v sw (2πf _o t) Then, the excitation signal of the mass-slave 20 is iterated as follows.

A first zero frequency excitation signal (n = 0)

e ₂ ° (t) = B ₀

is applied to the ground-slave 20. The coefficient B ₀ is then adjusted to an optimum value B ₀ 'such that the corresponding coefficient C ₀ of the error signal is minimum.

A new excitation signal is then applied to the ground-slave 20:

and the coefficient B ₁ and the phase% are adjusted so as to minimize the coefficient C ₁ of the error signal. Be B \ e \. φ \ the corresponding values. At the rank of iteration n, an excitation signal e ₂ "(t) given by:

(2.f _o ) t + φ ' ₂ ) + ... + B _n .sin (2π (nf ₀ ) t + φ _n )

Again, the coefficient B _n and the phase φ _B are adjusted to minimize the coefficient C _B.

The iteration continues in this way until the last rank n = N-I. Finally, the optimal signal of excitation of the mobile-slave mass

Is equal to: B ' _n .sm (2π (nf ₀ ) t + φ' _B )

It should be noted that this procedure can be performed at any time, even when the satellite is in flight.

Claims

A method of balancing the movement of the magnetic moving masses (10, 20) of a bilinear electrodynamic motor having two moving masses (10, 20) moving in opposite directions parallel to the axis (xx) of the engine, characterized in that said method comprises steps of:

providing at least one first magnetic sensor (12) and at least one second magnetic sensor (22) capable respectively of delivering a first electrical signal (^ ₁ (O) and a second electrical signal (s ₂ (t)) respectively representative of the movement of a first (10 and a second (20) moving masses,

- registering an error signal (Δs (t)) equal to the difference between said first (S ₁ (O) and second (s ₂ (t)) electrical signals and performing a harmonic analysis of said error signal, - applying at the first mobile mass (10) a sinusoidal excitation signal (e _λ (t)) at a given frequency / ₀ ,

- iteratively apply to the second moving mass N successive excitation signals (e ₂ "(t)), an excitation signal of rank n (O ≤ n ≤ N-ï) being equal to a series of order Fourier n of fundamental frequency / ₀ , the n-order term of said series being adjusted in amplitude and in phase so as to minimize the component at the frequency n / _{0 of} said error signal,

- Take as excitation signal of the second mobile mass excitation signal (^ ^" '(t)) obtained at the iteration of rank N-I.

The method of claim 1, wherein said magnetic sensors are Hall effect sensors (12, 22).

3. Application of the method according to one of claims 1 or 2 to the reduction of vibrations of bilinear electrodynamic motors.